Objective To explore the application of the automatic bone age assessment of BoneXpert software in Children and adolescent of China (Shanghai).Methods Left hand and wrist radiographs of 434 cases of children from trauma (264 boys and 170 girls) were obtained from Jan.2010 to Dec.2013.Bone age was assessed by BoneXpert software and 40 radiographs were randomly chosen and re-evaluated after a week.Results Seven images were rejected by BoneXpert,so the efficiency of the software was 98.4％ and it takes only 14.5 s for every image.Sample assessments by two times are exactly the same.As to BoneXpert software,there were statistical differences between bone age and chronological age in 8 groups:7,8,9,11 years old boys and 2,3,9,11 years old girls(P ＜ 0.05) out of 26 groups,and the average difference was about 1 s.The following modifications suggested that 0.57,0.79,0.93,-0.38 and-0.22,0.30,0.37,0.59 years could be added respectively for the groups of boys aged 7,8,9,11 and girls aged 2,3,9,11 years old according to their average differences between the bone age and the chronological age.Conclusions The BoneXpert method for automated determination of skeletal maturity which evaluates bone age objectively and quickly can be applied clinically after slight adjustment in some age groups.BoneXpert software is 100 percent automatic and could eliminate the subjective evaluation inaccuracy.%目的 探讨网络自动骨龄评估软件BoneXpert对中国(上海)儿童及青少年的适用性.方法 回顾性收集2010年1月至2013年12月因左手外伤在上海市儿童医院拍摄左手腕部X线片患儿的影像学资料.选择其中可用于人工骨龄评估的434张X线片,男童264张,女童170张.由自动骨龄评估软件BoneXpert评估骨龄.1周后简单随机法随机抽取40张X线片重新评估.结果 在434张X线片中,共有7张被BoneXpert评估软件拒绝,接受率为98.4％.每张评估时间约为14.5 s.2次抽样评估结果完全一致.在男女共26组中有8组(32％)骨龄评估结果与实际年龄间差异有统计学意义(P＜0.05),分别为男童7、8、9、11岁组及女童2、3、9、11岁组,但差异均值均在1 s左右.若按各年龄组骨龄评估结果与实际年龄间差异的均值进行相应调整,可将上述各组骨龄评估值分别上调0.57、0.79、0.93、-0.38岁及-0.22、0.30、0.37、0.59岁.结论 BoneXpert软件对左手腕X线片接受度高,诊断迅捷,评估过程100％自动化,完全消除了人工评估过程中的主观差异性.在大部分年龄段均适用于我国儿童,对少部分年龄段评估结果进行微调后可广泛应用于临床.
展开▼
